prikaz prve stranice dokumenta Primjena surogatnih modela za simulaciju kod numeričke optimizacije
Access restricted to authenticated users
master's thesis
Primjena surogatnih modela za simulaciju kod numeričke optimizacije
2021. urn:nbn:hr:179:420349
Marinović, Matko
University of Split
Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture
Department of Mechanical Engineering and Naval Architecture

Cite this document

Marinović, M. (2021). Primjena surogatnih modela za simulaciju kod numeričke optimizacije (Master's thesis). Split: University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture. Retrieved from https://urn.nsk.hr/urn:nbn:hr:179:420349

Marinović, Matko. "Primjena surogatnih modela za simulaciju kod numeričke optimizacije." Master's thesis, University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, 2021. https://urn.nsk.hr/urn:nbn:hr:179:420349

Marinović, Matko. "Primjena surogatnih modela za simulaciju kod numeričke optimizacije." Master's thesis, University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, 2021. https://urn.nsk.hr/urn:nbn:hr:179:420349

Marinović, M. (2021). 'Primjena surogatnih modela za simulaciju kod numeričke optimizacije', Master's thesis, University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, accessed 13 February 2025, https://urn.nsk.hr/urn:nbn:hr:179:420349

Marinović M. Primjena surogatnih modela za simulaciju kod numeričke optimizacije [Master's thesis]. Split: University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture; 2021 [cited 2025 February 13] Available at: https://urn.nsk.hr/urn:nbn:hr:179:420349

M. Marinović, "Primjena surogatnih modela za simulaciju kod numeričke optimizacije", Master's thesis, University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, Split, 2021. Available at: https://urn.nsk.hr/urn:nbn:hr:179:420349

Please login to the repository to save this object to your list.
Metadata
TitlePrimjena surogatnih modela za simulaciju kod numeričke optimizacije
AuthorMatko Marinović
MentorDamir Vučina (mentor)
GranterUniversity of Split
Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture
(Department of Mechanical Engineering and Naval Architecture)
Split
Defense date and country2021-09-23, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Mechanical Engineering
Abstract
U ovom radu su obrađeni surogatni modeli u svrhu numeričke optimizacije. Modelirana su četiri tipa surogatnih modela: neuralna mreža, model polinomne odzivne plohe, model radijalnih baznih funkcija i Kriging model, te je objašnjenja njihova teoretska pozadina. Modeli su primijenjeni na demonstracijskom primjeru optimizacije oblika stijenke dvodimenzionalnog simetričnog difuzora s dvije varijable optimiranja, s ograničenjima i bez ograničenja. Objašnjeni su postupci parametrizacije difuzora, definiranja ograničenja i izrade modela za numeričku simulaciju pomoću alata za računalnu dinamiku fluida. Opisane su temeljne značajke računalne dinamike fluida, metoda konačnih elemenata i diskretizacijske mreže. Objašnjene su osnove metoda optimizacije kod problema s ograničenjima. Modeliranje surogatnih modela je napravljeno unutar programskog paketa ModeFRONTIER. Planiranjem pokusa generirane su točke za ispitivanje domene optimizacijskog problema. Objašnjeni su korišteni parametri i postavke za modeliranje pojedinih surogatnih modela. Točnost svakog surogatnog modela je definirana i međusobno uspoređena na problemu sa i bez ograničenja. Na kraju su opisane metode uporabe surogatnih modela na složenijim inženjerskim problemima.
Abstract (english)
This thesis provides an overview on surrogate models in the field of numerical optimization. Four types of surrogate models were constructed: neural network, polynomial response surface, radial basis function model and Kriging model, also their theoretical background was explained. The models were used on a demonstration example of the wall shape optimization of a twodimensional symmetric diffuser using two optimization variables with constrains and without constraints. An insight was given into the parametrization of the diffuser, defining the constraints and the creation of a numerical simulation model using computational fluid dynamics tools. The basics of computational fluid dynamics, finite element method and finite element mesh were shown. The basic idea of constrained optimization was explained. The modelling of the surrogate models was done in the software ModeFRONTIER. Design of experiments technique was used to generate the sample points which were used to analyse the domain of the optimization problem. The used parameters and settings used in the modelling of each surrogate model were explained. The accuracy of each surrogate model was defined and compared on a constrained and unconstrained problem. The use of surrogate models on more complex engineering problems were discussed.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:179:420349
Study programmeTitle: Mechanical Engineering; specializations in: Structures and Energy Technology, Computer Aided Design and Engineering, Production Mechanical Engineering
Course: Computer Aided Design and Engineering
Study programme type: university
Study level: graduate
Academic / professional title: Magistar/magistra inženjer/inženjerka strojarstva (Magistar/magistra inženjer/inženjerka strojarstva)
Type of resourceText
File originBorn digital
Access conditionsAccess restricted to authenticated users
Terms of useLicence In copyright icon
Created on2022-11-04 13:34:13